Family and Generic Differentiation. 333 



and the Oyster-catcliers. From the original centre of 

 distribution of those proj^enitors, the two races spread and 

 radiated to occupy their present areas of distribution. They 

 carried with them the legacy of the ancestral germ-cells. 



How little environment has hafl to do with their colour- 

 pattern, is indicated by the additional fact, that not only is 

 the colour-])attern of the two nestling Stone-Curlews, from 

 the two countries already mentioned, identical in detail, but 

 their characteristic stripes and markings are also, for all 

 practical purposes, identical with the like markings on the 

 two Oyster-catchers already referred to ; a fact which may 

 and probably does indicate a community of descent or 

 phylogeny, but certainly cannot be interpreted in the sense 

 that the environmental details proper to these two forms 

 are identical. 



Returning however, to the more particular subject of the 

 actual correlation of colour-pattern with geographical or 

 faunal areas, many instances might be quoted. Mr. Witmer 

 Stone has given, for example, the case of the three genera of 

 eastern Cuckoos mentioned above ; but I must limit myself 

 here to one other example. 



In the genus Emberiza we have a group of species 

 {E.Jiaviventris, jioliopleura, major, aud cabanisi) from various 

 African localities which have a distinctive colour-pattern, 

 in which the factor of yellow is very prominent. In east 

 Siberia, Manchuria, and Japan, we find two species of 

 Emberiza {E. elegans and chrysophrys) in which although 

 yellow plays a part, the colour-pattern is distinctly differ- 

 entiated from the above. Both these groups are in turn 

 differentiated from another group comprising such species 

 as E. cia, cioides, castaneiceps, ciopsis, and stracheyi. My 

 contention is, that the colour-pattern characteristic of such 

 groups, not to mention others, must have a distinctive 

 genetic import, which is independent of environmental 

 effects and for which we ought, incidentally, to have some 

 method of expression in our system of classification.' We 

 must in classification, work our way down to such natural 

 compound units. (In this connection, see also remarks under 



